[Sussex] Munich picks its Linux distro

Geoffrey J. Teale gteale at cmedltd.com
Mon Apr 18 08:54:38 UTC 2005


Steve Dobson <steve at dobson.org> writes:

> Morin'

Morning..

>
> Well having read the article it said they were "migration from Microsoft
> Windows NT to Linux on the desktop".  This suggests that they haven't got
> the latest and greatest in terms of hardware.

Yup.  I've read that a lot of large organisations that have moved to
linux have been in this position.  If you're moving from NT4 then the
amount of work moving to Linux is actually comparable to "upgrading" to
Windows XP and the cost is much lower if you're not having to fork out
thousands on new hardware.  If you're looking for a "like for like"
replacement for NT4 era functionality then you don't need to be using
the latest and greatest Linux software, and the appeal of Debian's
stability and long release cycles starts to make a lot of sense.  It's
no coincidence that Novell is talking about a 12-24 month release
cycle for NLD and a much shorter one for SuSE Linux (I believe Red
Hat's Enterprise Linux has a longer release cycle than Fedora as
well).  Large organisations don't want to be upgrading every couple of
months, they want to know what they are dealing with.

If you live a little more on the cutting edge of technology that's
when the long release cycle starts to cause issues.

> I doubt very much that Peter Hofmann (the migration project leader) wanted
> Debian (or Debian-like).  The project was put out to tender and "received a
> large number of high-quality responses" so there is a good chance that 
> Ubuntu was in there somewhere, as was IBM.

I think there's a lot more politics going on.  If SuSE had still been
German owned I imagine they would have gone with the IBM/SuSE combo
they tried out.

> I thinks this speaks volumes for the F/OSS business model. Gonicus and
> Softcon, who I assume are local computer (at some level) consultants,
> were able to compete and beat the likes of IBM and HP based upon a 
> volunteer based disto.  The F/OSS model is allowing more diversity and
> therefore more customer choice.

Mmm, you're write that the F/OSS model allowed this kind of
competition, but I image the "local" part was important.  As I said,
had SuSE still been a German company instead of being a part of
an American company then the result would probably have been different.

> I also think is shows a mistake in Microsoft's approach.  Steve Ballmer's
> interrupted ski holiday to visit Munich's mayor was referenced.  M$
> sent their head man to talk to Munich Council's head person presumably 
> on the basis that where the head leads the body follows.
>
> While I don't know how German local government works, I suspect it is
> broadly like our own.  The mayor, while head of the council, is only one
> voice.  The mayor preferences don't count unless the council is tidy where
> the mayor's vote is the deciding vote.

You're also discounting that the German's as a nation have never
really been very pro Microsoft.  As well as recently being very pro
Linux they've historically been the strongest hold out of support for
OS/2 and BeOS.  In the latter case a German company even bought the
rights for BeOS from Palm and are now shipping a new version under the
title "Yellowtab Zeta OS"

> Well it isn't that difficult.  For any given manufacturer there are going
> to be a (very) limited set of options.

Indeed, but the difference between IBM (and Dell for instance) is that
IBM guarantee this, document it and stick to it.  We've been burned by
Dell making "promises" that they have failed to keep so I might have a
bee in my bonnet about this, but it's an issue that is part of an
attitude to the support of large organisations that clearly
distinguishes IBM from some of it's competitors (I can only hope that
the Chinese lot maintain this level of service).

> All of this allows IBM to manage the software problem to support their 
> product range.

Yup.

> Which adds to my point that the F/OSS business model is a very good one
> (I may even dare to say the best).  Gentoo is providing a solution to those
> that want to compile the latest, bleeding edge software.  

Actually, this is only a popular perception.  Gentoo by no means
provides the "latest" version of most packages.  It's just a reasonably modern version of the
software in most cases.

> It comes with
> its own limitations - installation being one of them.  And before I get
> flamed I am NOT saying that Debian doesn't have limitations, just that it
> has different limitations.  When picking a distro pick the one who's 
> limitations provide the least problems to the way YOU want to do thinks.

Absolutely, probably the most balanced and sensible comment about
distro's I've seen here for a long time.

Contrary to popular belief I'm not anti-Debian, I'm anti the "you
_must_ use this distro" approach.  Luckily its become a running joke
on this list, but the Debian, Gentoo, <name any distro here>
communities are often filled with zealots, I get very angry about that
and often fight back.  I've seen no end of unfair criticisms of Gentoo
in the last few years (and often lashed back with accusations of my
own) but what it all boils down to is this:  not every distro is the
same and the world would be a poorer place if it were.  A uni-distro
world would be almost as bad as a world with just one OS.  

I'm interested in all distro's and seeing where they fit best, I try
everything I sensibly can and return to the ones that seem best to
achieve what I need to in the short term.  The only machine in my
network that is fixed on one thing for any length of time is the Sun
Ultra5 that runs my website and a few other "key" services.  This runs
OpenBSD, not because I particularly like OpenBSD, but it works well on
that box and the Apache in a chroot environment gives me warm fuzzies
:-)

-- 
Geoff Teale
CMed Technology            -   gteale at cmedresearch.com
Free Software Foundation   -   tealeg at member.fsf.org




More information about the Sussex mailing list